Distortion mod for dyacomp/Ross

[analogmike]

HI,

Due to a soldering error we found an easy way to get distortion from a dynacomp or Ross. Just disconnect the ground from the 10uF cap that connects to the three transistors. It's not going to be good enough to sell your Big Muffs etc but it's cool in a lo-fi way. You can put a toggle there, and the SUSTAIN knob acts as a distortion control. I do not guarantee that the op amp won't be harmed by this, you are modifying the control voltage to a level that it may not be meant to see. Circuit bending can make cool sounds but not always safe.

have fun!

[Mark Hammer]

Envelope ripple for fun and profit, eh?

To the uninitiated, that 10uf cap is what normally smooths out the envelope signal enough that you DON'T hear every blip in the envelope and get a sort of "smooth tachometer reading" of the input signal.

When some folks complain about "distortion" in their envelope-controlled filters (autowah), very often that sonic quality or "Scottish burr" on the decay phase is a result of the smoothing/averaging cap being too small for their particular playing style.  Mike's mod simply says" Okay, if too little capacitance gets you a burr, then NO capacitance/smoothing ought to get you buzz.  And clearly it does!

Nice thinking outside the box!

Now it kinda makes one wonder, what would happen if you omitted the smoothing cap from, say, a Dr Q or a Mutron?

[Barcode80]

how did it sound mike?

[analogmike]

It sounds like a cheap distortion pedal with low battery, might be useful for some types of sounds. It's basically a free mod so I thought someone might like to try it.

have fun and thanks for explaining it better, Mark!

[Jay Doyle]

...I do not guarantee that the op amp won't be harmed by this, you are modifying the control voltage to a level that it may not be meant to see. Circuit bending can make cool sounds but not always safe.

You are fine, the 27k resistor leading to the Iabc port on the 3080 limits the current into the chip to safe levels no matter what voltage the sustain pot feeds it.

[kvb]

So it still sustains?   but just all rippled?

This would still also be sensitive to picking attack/ level; soft picking = less ripple/ hard picking = more ripple?

[analogmike]

Thanks Jay!

Not much compression/sustain after this mod, just distortion. The control circuit which makes the op-amp compress is now out of control when you lift that cap.

[Jay Doyle]

...I do not guarantee that the op amp won't be harmed by this, you are modifying the control voltage to a level that it may not be meant to see. Circuit bending can make cool sounds but not always safe.

You are fine, the 27k resistor leading to the Iabc port on the 3080 limits the current into the chip to safe levels no matter what voltage the sustain pot feeds it.

I realized after writing this that that statement could be misunderstood to mean that the chip would be safe at all power levels. But the 27k is only applicable to supply voltages of 9V or less. If you try to pump your comp at a higher voltage than 9V, you are playing with fire or you need to up the 27k resistor to limit the amount of current into the Iabc port to 2mA.

[cleanroom]

I finally got around to doing this mod.  The sound is pretty cool.  It appears that the Comp knob is gone.  Any thoughts on other controls that could be added to this or at least a good place to start tinkering?  I was toying with the idea of adding a footswitch to go between the distortion and comp. 
Thanks Mike for the mod..

[analogmike]

Glad you tried it!!

I was recently wondering if this could get close to the Clay Jones overdrive which some people claim is based on a messed up compressor...

But I'm sure he did a lot of tweeking to get it to sound good. Anyone can make a good overdrive with 10 knobs and 5 toggle switches, tuning it to work with just a few knobs is the tough part.

[petemoore]

  The resistance introduced between the cap and where the top of the three resistors? Something like a 1 meg pot for try outs?

[cleanroom]

I wonder how playing with the diodes would effect this?  I didn't care for Ger diodes when this was a comp. I thought it clamped down too early.  Would this perhaps give you more distortion with the lower voltage?  hmm...

[Mark Hammer]

I finally got around to doing this mod.  The sound is pretty cool.  It appears that the Comp knob is gone.  Any thoughts on other controls that could be added to this or at least a good place to start tinkering?  I was toying with the idea of adding a footswitch to go between the distortion and comp. 
Thanks Mike for the mod..

Well, if you want to get real far gone, try lifting one of the collectors from Q3 or Q4 (doesn't matter which).  Mike's mod essentially removes all the envelope smoothing from the rectifier so that the 3080 is modulated by the raw rectified signal.  The design uses a full-wave rectifier, so that if you disconnect one of the transisitors from the node where the two half-waves are summed (where the two collectors, the 150k resistor, and 10uf cap to ground meet), you are essentially left with modulation by one half cycle, which will be at half the frequency of the full-wave version.  This modulation should be far more audible and may even yield something that sounds like octave doubling.

Remember that if you feed a 4-quadrant multplier with the same signal at its audio input and its modulation input, you end up with double the frequency,  Since using only one half-cycle will not be exactly like the input, it won't be exactly like the octave-doubling outlined, but it will be somewhere in the neighbourhood, and may well have an interesting sound.  I've got a recently finished Ross clone on the bench.  I think I'll try it out tonight.

So, just to recap: lift the ground side of the 10uf cap, lift the collector of one of the transistors in the rectifier circuit....and stand back.

[kvb]

so what will be happening then?

The axe signal is not rectified, but it is affected (modulated?) by a rectified version of itself?

[Mark Hammer]

So I did the mod last night, as described.  As Mike indicates, you do get a distorted sound.  And as I predicted, you get something approximating an octave fuzz...but only approximating.  This is NOT a cheap substitute for a Foxx or Tychobrahe.  Interestingly, and as one might predict from theory, the sound gets MORE distorted and octave-ey as it decays, the reason being that ripple generally becomes more noticeable during the decay phase when one uses half-wave rectification.  Initially, you have the "folded-over" peaks of all those harmonics, but as the sound simmers down to primarily the fundamental and lower-order harmonics, the gaps between those successive peaks are not filled up with anything.  In the absence of a smoothing/averaging cap, they are seriously not filled up with anything, such that the 3080 starts to become modulated at the rate of those peaks and gaps.  As the string settles down to primarily the fundamental frequency, you have that state where the VCA is being modulated at the same rate as the audio input signal, and you start to get frequency doubling.  Not quite as predictable or rich a transformation as the Gretsch Controfuzz (which also starts out cleanish and gets fuzzier over time), but interesting.

Happily, the Tonepad layout provides an easy means to install this mod.  Cut the copper trace where you see the yellow line.  Drill two small holes where you see the yellow dots.  Link these two points via a SPST toggle or other type of switch.  When the switch is open, the 10uf cap and one of the collectors will be lifted from the common tie-point.  When you close it, you're back to a perfectly normal Dynacomp or Ross.
Personally, I don't think I'll keep it on mine since I have more than enough ways to get distortion at my disposal, but for those who find this is a  useful quick and dirty something-for-nearly-nothing mod, I tip my hat to Analog Mike for sharing his happy accident.

so what will be happening then?

The axe signal is not rectified, but it is affected (modulated?) by a rectified version of itself?

The axe signal IS rectified, but poorly.  That rectified signal is fed to the 3080 and the signal is modulated by itself.  Normally, with nicely smoothed full-wave rectification, the 3080 is modulated by the overall guitar signal.  As you systematically remove those things that produce a smooth rectified signal, that modulation starts to get more "granular".  It's a bit like the difference between having a conversation based on the idea of what someone says overall vs them spelling out each word really fast.  In this case, they're spelling out every second letter.

[cleanroom]

Thanks for the follow up Mark.

I was curious how the lifting of the collector would turn out.  How different was the sound between the lifting of th 10uF cap and the lifting of both the 10uF cap and the collector?

[Jay Doyle]

Well, for an additional option, while we are tearing apart the circuit, you can add a DPDT to switch the FWR input between the 'stock' which is in series with the OTA, to lifting the caps feeding the FWR and moving their inputs to the input buffer transistor's collector and emitter. This then runs the comp in parallel mode, which does produce a different response, no idea how it would work in this distortion sense, but as a regular comp, it does change the sound.

Jay

[Mark Hammer]

I didn't do both.  My bad.  I went after the most robust and easiest-to-implement mod.  This one just involved taking my un-boxed Ross board and making one small slice with my X-acto.  As it happened, that one slice lifted the cap and one collector out of the path.  Leaving the two halves of the rectifier in place, and simply lifting the cap will undoubtedly result in less audible distortion, simply because the modulation rate will be essentially doubled.  Of course, that may be the more desirable of the two potential distorted sounds obtainable via this route.  I'll leave that to others to find out for themselves.

Incidentally, since the cap is out of circuit, variations in the "recovery" resistor (the 150k unit that often gets called "Attack") should not have any impact, since it determines charge-up time for the cap, and without a cap to charge up....well.

[petemoore]

  Incidentally, since the cap is out of circuit, variations in the "recovery" resistor (the 150k unit that often gets called "Attack") should not have any impact, since it determines charge-up time for the cap, and without a cap to charge up....well.
  I happened to do some mods between Q3/Q4, the 10uf cap became 3u3, the 150k resistor became 150k - 1.15M, I added the 1m pot.
  The cap/R mod started as 3u3/480k, and did about what I expected, making it a more unruly, harder compression, deeper attenuation and longer recovery time.
  The 1m pot seems to make a little difference up to...1/3 rotate, or I'll say 480k, above that estimated resistance, no noticable difference. But there is a difference between 150k and up, this can be heard as the circuit is settled on a sustaining note or chord, by slight volume rise as knob turns, maybe a slightly different compression rate.

[Jay Doyle]

A quick list off the top of my head on how to bend the Ross/Dyna circuit to distort:

Overdrive the 3080 with a hot signal input. Ways to do this: Increase the value of the trim to reduce it's loading effect on the input buffer; change the input transistor from an NPN buffer to a PNP booster; run a booster in front of the comp; any combination.

Clip the 3080's output. To do this the easiest, simply keep one end of the 27k resistor on the Iabc port and switch the other end's connection from the Sustain pot's wiper to a direct connection with V+, this will put the 3080's gain at max; you can also increase the 150k resistor after the OTA (this option with significantly increase the noise).

Use the FWR as the output. Take the 10uF out of the circuit and take the output from that point and you have a rough freqency doubler; use that signal to directly drive the circuit instead of integrated in the cap (this is what is being done above)

Drive the OTA's Iabc port with the original signal: this is a rough four quadrant multiplier, input the buffered signal from the input transistor straight to the Iabc port.

I'm sure there are many more. Overdriving the input is worth looking into if you have the chance and a spare 3080.

Jay Doyle